When advertisers purchase distribution for their ads on a web site or set of web sites, there are a few standard ways they can pay for that. The simplest models are:
1. A fixed cost per thousand impressions (“CPM model”). If an advertiser pays a fixed rate CPM (e.g., $1.00), they pay the fixed rate of $1 for every thousand times their ad is shown (impression), regardless of user response rate to that ad.
2. A fixed cost-per-click model (“CPC”). If an advertiser pays a fixed rate (e.g., $1) CPC, they pay $1 every time a user clicks on that ad, regardless of the number of times the ad is shown without being clicked.
However in many situations the market value of ad space may vary. For example, if the ad is to be matched to a keyword on an active search page or in contextual advertising, then the value of that ad space depends on the market value for that keyword (e.g., “casino” is typically more valuable for advertising than “paper”).
As a result, auction models have become common. In this case advertisers bid on how much they're willing to pay—typically on a CPC basis—to have their ad shown on those pages, against that keyword or context, etc. An ad server implements the placement algorithm and is able to maximize the value of that ad space by selecting the highest paying ads at any given time. In some cases the ad server will also combine performance data for that ad (including click-through-rate (“CTR”) data, for example) with the bid price per click to determine the effective CPM (“eCPM”) rate for each ad, and then choose the highest eCPM ads. Or combining with purchase or other conversion data to establish a cost per action (“CPA”), and then include CPA values among the selection process. In either case the ad selection formula typically relies on an auction-based marketplace. The term eCPM is an industry standard known to persons of ordinary skill in the art. As is readily understood by a person of ordinary skill in the art, CPA is also known as cost-per-conversion, or cost-per-sale.
The eCPM value reflects what the equivalent CPM is if the pricing model is based on CPC or some other non-CPM model. For example, a CPC rate multiplied by the ad's click-through-rate multiplied by 1000 yields the eCPM for that ad based on its response.
                    eC        ⁢                                  ⁢        P        ⁢                                  ⁢        M            =              C        ⁢                                  ⁢        P        ⁢                                  ⁢        C        *        C        ⁢                                  ⁢        T        ⁢                                  ⁢        R        *        1        ,        000              ;    or            eC      ⁢                          ⁢      P      ⁢                          ⁢      M        =                  Cost        Click            *              Click        Impression            ×      1      ,      000      
Auctions provide a means for extracting appropriate market value for ad space, but they also create problems. For example, when advertisers purchase a fixed CPM or CPC campaign to distribute their ads with a publisher or media company, they know that they will get that distribution or otherwise there was an issue with the vendor. In an auction marketplace the advertiser does not have the same clear contract with the publisher, since other advertisers may outbid them for the distribution at any time. As an example, if the advertiser wants to run a campaign that generates 30,000 clicks in a month within a budget of $15,000, then they might bid a maximum CPC of $0.50 per click, they might start off getting the 1000 clicks per day that they want, then a few days later suddenly drop to 100 or 0 clicks per day when their ad is preempted by a higher paying advertiser. At that point they have to adjust their bid to a higher per click rate to restart the campaign (risking exceeding the original advertising budget), or move the campaign to another publisher, etc.
As such, fixed price models are good for advertiser predictability but bad for publishers and networks looking to maximize the value of dynamic properties. Auction models are good for publishers and networks' ability to maximize value, but bad for advertiser predictability. Advertisers generally prefer to pay based on a CPC rate, thus assuring they pay only when users show interest in their ad and generate a response. Inapposite, publishers like to be paid on a CPM basis, thus providing a more predictable return—publishers know how much traffic they get to their site (how many pages they serve to their users per-day or per-month); so on a CPM basis they can predict revenue independent of an advertising campaign performance.
Furthermore, the ad traffic generated by users visiting publishers' sites can be anywhere from a very low value to a very high value for advertising. Even within a single publisher the value of traffic generated often has a range of value to advertisers. However, without an ability to classify that traffic into different groups that separate the higher from the lower value traffic, publishers typically must strike simplistic CPM deals with the networks that deliver advertisements. These deals provide publishers with a flat CPM rate that does not give publishers upside on their higher value traffic.
Further, missing from the art are methods and systems for placement of internet advertisements or services that provide publishers an upside value on their higher value traffic, and provide publishers with a more predictable return. The present invention can satisfy one or more of these and other needs.